Efficacy and safety of high-dose pravastatin in hypercholesterolemic patients with well-compensated chronic liver disease: Results of a prospective, randomized, double-blind, placebo-controlled, multicenter trial


  • This article was presented in part at the 57th annual meeting of the American Association for the Study of Liver Diseases, October 31, 2006, Boston, MA. Clinical trial registration number NCT00529178, available at: http://clinicaltrials.gov

  • Potential conflict of interest: Dr. Lewis is a consultant for Bristol-Myers Smith. Dr. Mortenson owns stock in and is an employee of Bristol-Myers Smith. Dr. Fusco owns stock in Astra and Bristol-Myers Smith. He is an employee of Johnson and Johnson. Dr. Belder owns stock in Bristol-Myers Squibb.


The hepatotoxic potential of 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors in patients with underlying chronic liver disease remains controversial. We performed a multicenter, randomized, double-blind, placebo-controlled, parallel-group trial that compared pravastatin (80 mg) to a placebo administered once daily to hypercholesterolemic subjects greater than 18 years of age with at least a 6-month history of compensated chronic liver disease and with a low-density lipoprotein cholesterol (LDL-C) level greater than or equal to 100 mg/dL and a triglyceride (TG) level lower than 400 mg/dL. The efficacy was determined by the percentage change in LDL-C [along with the total cholesterol (TC), high-density lipoprotein cholesterol, and TG] from the baseline to week 12. The safety was analyzed by the proportion of subjects who developed at least 1 alanine aminotransferase (ALT) value greater than or equal to 2 times the upper limit of normal for those with normal ALT at the baseline or a doubling of the baseline ALT for those with elevated ALT at the baseline during 36 weeks of treatment. A total of 630 subjects were screened, and 326 subjects were randomized; nonalcoholic fatty liver disease was present in 64%, and chronic hepatitis C was present in 23%. In the intent-to-treat population, pravastatin (80 mg/day) significantly lowered the mean LDL-C, TC, and TG values at week 12 and at other times (weeks 4, 8, 24, and 36) in comparison with the placebo. The incidence of subjects who met the primary prespecified ALT event definition was lower in the pravastatin group at all times over the 36 weeks of therapy in comparison with the placebo group, although the difference was not statistically significant. No differences were seen on the basis of the baseline ALT values or among the different liver disease groups. Conclusion: High-dose pravastatin (80 mg/day) administered to hypercholesterolemic subjects with chronic liver disease significantly lowered LDL-C, TC, and TGs in comparison with the placebo and was safe and well tolerated. The concern over an increased potential for statin-induced hepatotoxicity in patients with chronic liver disease appears to be lessened on the basis of these results. (HEPATOLOGY 2007.)

Pravastatin sodium is a 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase inhibitor that acts as a cholesterol-lowering agent by principally reducing low-density lipoprotein cholesterol (LDL-C) levels.1, 2 HMG CoA reductase inhibitors have been associated with abnormalities of liver-associated enzymes (LAEs),3 which are generally mild, dose-related, nonprogressive elevations in serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT). On the basis of animal studies, the elevations may be the result of mevalonic acid depletion during the inhibition of cholesterol synthesis.4 A number of meta-analyses and studies of pooled data from randomized, placebo-controlled trials of statins used for the treatment of hyperlipidemia or for the primary or secondary prevention of cardiovascular disease have assessed the risk of liver function test (LFT) abnormalities in patients without chronic liver disease.5–7 The proportion of patients having ALT or AST abnormalities is low, averaging only about 1%. In placebo-controlled clinical trials specifically involving pravastatin, abnormal levels of ALT and AST from 3 long-term studies involving 19,592 subjects occurred with a similarly low frequency (≤1.2%) in both treatment groups. In addition, no statistically significant between-group differences in LFT abnormalities were observed in subjects with mildly elevated ALT values [up to 3 times the upper limit of normal (ULN)] at the baseline in these pooled trials.7

Although isolated reports of symptomatic hepatic injury with statins have appeared since lovastatin was first introduced,3, 6, 8–13 instances of severe injury, including acute liver failure, have been very rare.3, 14 Nevertheless, despite the favorable risk-benefit profile of pravastatin and related agents with respect to hepatic safety6, 15–18 and the call for more aggressive lipid lowering to reduce the risk of cardiovascular disease,19 the use of statins remains a labeled contraindication in patients with underlying liver disease.20 As a result, concern for statin-related hepatic injury remains pervasive in clinical practice and is a frequent source of consternation and consultation for gastroenterologists and hepatologists.21

In an effort to examine the potential hepatotoxic risk of statins in patients with underlying liver disease, a number of retrospective studies have been performed in individuals with elevated ALT levels. Tolman9 and Chalasani and colleagues22–24 have demonstrated the relatively safety of lovastatin, atorvastatin, and simvastatin in comparison with a placebo in individuals with predominantly nonalcoholic fatty liver disease (NAFLD). More recently, the safety of statins was reported in cohorts of hepatitis C patients.25, 26 In addition, open-label pilot studies suggest that pravastatin and atorvastatin can be given safely for the treatment of hypercholesterolemia in patients with NAFLD and primary biliary cirrhosis (PBC) and in the post–liver transplant setting.27–30 The National Lipid Association's Statin Safety Task Force hepatology panel31 recommended that statins could be used in appropriate patients with NAFLD, although they recognized that this recommendation is at odds with current prescribing information.21 At present, there continue to be concerns raised by clinicians and regulatory bodies alike about the use of statins in patients with liver disease.16, 32, 33

To date, however, no randomized controlled clinical trial has evaluated the effect of any statin in patients with well-defined chronic liver disease. The purpose of the current trial was to prospectively determine the efficacy and safety of high-dose pravastatin in hypercholesterolemic subjects with a variety of well-compensated chronic liver diseases. This study had its roots in the over-the-counter development program that was being planned for pravastatin in order to satisfy US Food and Drug Administration concerns about the use of pravastatin in individuals who might have one of several unsuspected forms of chronic liver disease. The maximum US Food and Drug Administration–approved dose of pravastatin (80 mg/day) was selected to ensure that we captured as many dose-related effects as possible. The safety and efficacy of the 80-mg daily dose has been demonstrated in a pooled analysis of 2 multicenter, double-blind, placebo-controlled studies of patients with primary hypercholesterolemia without liver disease, with the safety and tolerability similar to those of the placebo.34


AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; HDL-C, high-density lipoprotein cholesterol; HMG CoA, 3-hydroxy-3-methylglutaryl coenzyme A; ITT, intent-to-treat; LAE, liver-associated enzyme; LDL-C, low-density lipoprotein cholesterol; LFT, liver function test; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease; PBC, primary biliary cirrhosis; SD, standard deviation; TC, total cholesterol; TG, triglyceride; ULN, upper limit of normal.

Patients and Methods

Study Objectives

This was a multicenter, randomized, double-blind, placebo-controlled, parallel-group trial comparing the efficacy and safety of pravastatin (80 mg) versus a placebo administered orally once daily in hypercholesterolemic subjects [mean LDL-C level ≥ 100 mg/dL and triglyceride (TG) level < 400 mg/dL] with chronic, well-compensated liver disease. The primary efficacy objective was to determine the percentage change from the baseline to week 12 in serum LDL-C. Secondary efficacy objectives were the percentage changes from the baseline to week 12 in high-density lipoprotein cholesterol (HDL-C), total cholesterol (TC), and TG by pravastatin (80 mg) versus the placebo. The primary safety objective was to estimate in the same population the proportion of subjects with at least 1 ALT value greater than or equal to 2 times the ULN) for those with normal ALT (≤ULN) at the baseline or a doubling of the baseline ALT for those with elevated ALT (>ULN) at the baseline over the 36 weeks of treatment.

The protocol and its amendments and the subject informed consent received institutional review board/independent ethics committee approval prior to the initiation of the study. Informed written consent was obtained from all subjects.

Study Design

The study was sponsored by Bristol-Myers Squibb, which developed the protocol with assistance from an external hepatic safety and steering committee. Eighty-one study sites were initiated, of which 69 enrolled subjects (64 in the United States and 5 in Mexico; see the appendix for the list of investigators). During the screening, the subjects were not permitted to take any lipid-lowering treatments and entered a 4-week, single-blind, placebo lead-in period, during which time they were to have adhered to the National Cholesterol Education Program Step 1 diet.35 At the end of the placebo lead-in period, the subjects who had mean serum LDL-C levels greater than or equal to 100 mg/dL were randomly assigned in a 1:1 ratio to receive 80 mg of pravastatin (as two 40-mg tablets) or a matching placebo during the double-blind, 36-week treatment period.

Inclusion Criteria.

Men and women, 18 years or older, who satisfied all of the following criteria were eligible to enroll in the study:

  • 1Those with any chronic, well-compensated stable liver disease defined as follows: chronic hepatitis C documented by a positive viral load test, nonalcoholic fatty liver (NAFL)/NAFLD documented by radiological imaging and/or liver biopsy, or any other chronic liver disease documented by the usual clinical parameters.
  • 2Those whose LDL-C was greater than 100 mg/dL and whose serum TGs were less than 400 mg/dL.

Exclusion Criteria.

Subjects were ineligible and were excluded from the study if they met any of the exclusion criteria:

  • A pregnant or lactating female or a woman of childbearing potential who was unwilling to use an adequate method of contraception during the study.

  • Any of these laboratory abnormalities: ALT or AST levels greater than 5 times the ULN, a total bilirubin level above the normal limit, a serum creatinine level greater than 1.5 mg/dL, a creatine kinase level greater than 3 times the ULN, an albumin level below the lower limit of normal, a prothrombin time greater than 2 seconds, and a platelet count below the lower limit of normal.

  • Ascites, jaundice, or cirrhosis with a Child-Pugh score greater than 5.

  • A disorder affecting serum bilirubin (for example, hemolytic anemia, Gilbert's disease, and sickle cell anemia).

  • Antiviral therapy for hepatitis B or C.

  • Prior lipid lowering medications or therapy for 8 weeks or more.

  • Cancer (other than basal cell carcinoma) or cancer chemotherapy.

  • Significant cardiovascular, cerebrovascular, renal, or thyroid disease or uncontrolled diabetes mellitus within 6 months prior to randomization.

The subjects were to have the study drug discontinued if any of the following study-specific criteria emerged:

  • 1ALT or AST values greater than or equal to 10 times the ULN if the value remained elevated upon repeat testing within 1 week.
  • 2Symptomatic liver disease, as defined by the development of severe malaise, ascites, jaundice associated with direct hyperbilirubinemia, or a Child-Pugh score greater than 5 in subjects with known cirrhosis.
  • 3An alteration in the hepatic synthetic function on 2 consecutive assessments, which was defined as follows: an albumin level lower than 2.8 mg/dL, a prothrombin time prolonged for 4 seconds or more, a total serum bilirubin level greater than 1.5 mg/dL with more than 50% direct bilirubin, or a creatine kinase level greater than 10 times the ULN.
  • 4Any reason based on the opinion of the investigator.

After the withdrawal of the study drug, all patients continued study participation, including regularly scheduled visits and laboratory assessments.

An independent data monitoring committee was established to evaluate the incidence of adverse events (AEs), laboratory measurements, and safety assessments to ensure the ongoing safety of the study subjects. A reporting statistician, not involved in the conduct of the study, provided the independent data monitoring committee with unblinded safety data during the study. The independent data monitoring committee reviewed safety data on an interim basis and made recommendations on the study continuation.

Statistical Methods

Data from all sites were combined for the purpose of analysis. All analyses were conducted with SAS version 8.2 or higher. All statistical testing was performed with 2-sided tests at the 5% significance level.

Sample Size and Power.

Randomization was executed in an attempt to balance the number of subjects enrolled in each treatment group by the type of liver disease. Assuming that 150 subjects would be randomly assigned to each treatment arm, a 2-sided significance level of 5%, and a standard deviation for the percentage change in LDL-C of 12%, this study could detect a difference in the mean percentage change in LDL-C between the treatment groups of 15% with greater than 99% power. The study was not specifically powered to address the safety endpoint (that is, the proportion of subjects who had an ALT value that doubled), as this would have required more than 3000 subjects. However, a study sample size of 150 per treatment group was considered adequate to provide approximately 20% power for an equivalence interval of 20% ± 10%.36

All efficacy analyses were performed on the intent-to-treat population, which was defined as all randomized subjects who took at least 1 dose of double-blind medication and had at least 1 postbaseline measurement. The safety population was defined as all subjects randomized who took at least 1 dose of the study drug.

Statistical Analyses.

The primary efficacy endpoint (the percentage change from the baseline to week 12 in LDL-C) was analyzed with a last observation carried forward algorithm for subjects who withdrew prematurely or missed an intermediate efficacy assessment. An analysis of covariance was used, and the primary comparison of interest was the treatment group of active pravastatin versus the placebo. Least square means and standard errors were obtained from the model with 95% confidence intervals for the differences of the least square means between the active treatment groups versus the placebo. In cases of severe nonnormality, nonparametric techniques were employed to analyze the data. The percentage changes from the baseline to week 12 in HDL-C, TC, and TG were the secondary endpoints, and they were analyzed with the same analysis of covariance model and summary statistics specified for the primary endpoint.

The primary safety endpoint was the overall proportion of subjects with at least 1 ALT value at least 2 times the ULN for those subjects with normal ALT (≤ULN) at the baseline or a doubling of the baseline ALT for those subjects with elevated ALT (>ULN) at the baseline (called an ALT event) within 36 weeks of once daily administration. This was considered the simplest and most direct means of detecting a potentially meaningful difference between the treatment groups, with the clinical significance of such a difference being addressed with other parameters, including associated symptoms and concomitant bilirubin elevations. Descriptive statistics of the baseline ALT values, ALT values at each scheduled visit, and changes in the ALT values from the baseline to each scheduled visit were assessed.



During the period between March 5, 2003 and July 23, 2004, 630 subjects were screened, and 446 were enrolled in the placebo lead-in phase. Of these, 326 subjects were randomly assigned to a treatment (163 to pravastatin and 163 to placebo). Overall, 133 (81.6%) of the pravastatin subjects completed the double-blind period versus 118 (72.4%) in the placebo group. The overall safety population was 163 in each group, with 160 in each group serving as the intention-to-treat population.

Disposition by the Baseline Liver Disease

Table 1 shows the distribution of liver disease diagnoses by the treatment group assignment. The profile for subjects with chronic hepatitis C and NAFL/NAFLD was similar to that observed for all subjects. As histology was not required for entry, the exact number of subjects with advanced fibrosis or cirrhosis could not be determined. Only a few subjects were known to have cryptogenic cirrhosis or PBC, as noted in the footnote to the other liver group in Table 1. However, no subject with Child B or C cirrhosis would have been allowed to enroll, given the exclusion criteria. Similarly, the collected information did not permit distinguishing subjects with simple hepatic steatosis from those with steatohepatitis.

Table 1. Summary of All Subjects by the Baseline Liver Disease
 Chronic Hepatitis CNAFL/NAFLDOther*
PravastatinPlaceboTotal (n = 126)PravastatinPlaceboTotal (n = 291)PravastatinPlaceboTotal (n = 54)
  • *

    The specific etiologies for pravastatin and placebo were hepatitis B (n = 2 and n = 1, respectively), hemochromatosis (n = 5 and n = 2, respectively), autoimmune hepatitis (n = 2 and n = 3, respectively), past alcohol use (n = 6 and n = 6, respectively), and cirrhosis due to PBC or cryptogenic cirrhosis (n = 5 and n = 4, respectively). Abbreviations: ITT, intent-to-treat; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease.

Screen failures9118
Completed placebo lead-in period8020835
Discontinued in lead-in period377211
Randomized [n (%)]38 (23.3%)43 (26.4%)81 (24.8%)105 (64.4%)104 (63.8%)209 (64.1%)20 (12.3%)16 (9.8%)36 (11%)
Completed double-blind period312859 (72.8%)8877165 (78.9%)141327 (75.0%)
Discontinued in double-blind period7 (18.4%)15 (34.9%)22 (27.2%)17 (16.2%)27 (26.0%)44 (21.1%)6 (30.0%)3 (18.8%)9 (25.0%)
Safety population38 (100%)43 (100%)81 (100%)105 (100%)104 (100%)209 (100%)20 (100%)16 (100%)36 (100%)
ITT population38 (100%)43 (100%)81 (100%)103 (98.1%)101 (97.1%)204 (97.6%)19 (95.0%)16 (100%)35 (97.2%)

Demographics and Patient Characteristics

The treatment groups of the randomized subjects were well balanced, and no statistically significant differences were observed. The majority of subjects in both groups were white, and the mean age was 50 years. Approximately one-third of all randomized subjects were Hispanic. The demographics for all randomized subjects are summarized in Table 2. Prior lipid-lowering drug use was uncommon, being found in only 14 randomized subjects overall (9 in the pravastatin group and 5 in the placebo group), with only 3 individuals receiving prior statin therapy (all in the placebo group).

Table 2. Demographic and Baseline Characteristics (All Randomized Subjects)
ParameterPravastatin (n = 163)Placebo (n = 163)Total (n = 326)
  1. SD indicates standard deviation.

Age (years)   
 Mean (SD)49.8 (10.18)49.9 (11.54)49.8 (10.86)
Gender [n (%)]   
 Male77 (47.2%)92 (56.4%)169 (51.8%)
 Female86 (52.8%)71 (43.6%)157 (48.2%)
Race [n (%)]   
 White140 (85.9%)151 (92.6%)291 (89.3%)
 African American13 (8.0%)3 (1.8%)16 (4.9%)
 Asian5 (3.1%)4 (2.5%)9 (2.8%)
 Other5 (3.1%)5 (3.1%)10 (3.1%)
Body mass index (kg/m2)   
 Mean (SD)30.605 (6.0457)31.305 (6.3027)30.956 (6.1763)

LAE Test Values.

All pre-enrollment LAE values were similar for both treatment groups (Table 3). The mean historical ALT and AST values were statistically significantly lower in the pravastatin group than in the placebo group (P = 0.0073) but were not considered clinically significant.

Table 3. Liver-Associated Test Values at the Baseline (All Randomized Subjects)
ParameterPravastatin (n = 163)Placebo (n = 163)P*
  • *

    The P values are based on a 2-sample t test. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; SD, standard deviation.

ALT (IU/L)  0.0316
 Mean (SD)65.7 (44.77)77.3 (49.01) 
AST (IU/L)  0.0073
 Mean (SD)41.8 (24.73)50.5 (31.51) 
Total bilirubin (mg/dL)  0.5648
 Mean (SD)0.6078 (0.33325)0.6290 (0.30938) 
Creatine kinase (IU/L)  0.2423
 Mean (SD)157.9 (377.25)122.5 (75.78) 

Baseline Lipid Levels: The Intent-To-Treat Population.

All baseline lipid values were similar for both treatment groups (Table 4). The mean HDL-C was modestly (but statistically significantly) higher at the baseline in the pravastatin group than in the placebo group (P = 0.0392).

Table 4. Baseline Lipid Levels (ITT Population)
ParameterPravastatin (n = 160)Placebo (n = 160)P*
  • *

    The P values are based on a 2-sample t test for continuous variables. Abbreviations: ITT, intent-to-treat; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SD, standard deviation; TC, total cholesterol; TG, triglyceride.

LDL-C (mg/dL)  0.6174
 Mean (SD)138.8 (26.74)140.5 (31.50) 
HDL-C (mg/dL)  0.0392
 Mean (SD)49.2 (13.71)46.4 (10.39) 
TC (mg/dL)  0.9738
 Mean (SD)219.0 (33.40)219.1 (37.66) 
TG (mg/dL)  0.5571
 Mean (SD)161.2 (78.51)166.2 (71.45) 

Efficacy Results

As expected, the treatment with pravastatin (80 mg/day) statistically significantly lowered the mean LDL-C, TC, and TG values at week 12 in comparison with the placebo (Table 5). At other times (weeks 4, 8, 12, 24, and 36), there was also a statistically significant reduction in LDL-C, TC, and TGs favoring the pravastatin treatment. The difference in HDL-C between groups was statistically significant, favoring pravastatin, only at weeks 4 and 8 (P = 0.005 and P = 0.003, respectively). No statistical differences were observed in any change from the baseline of the lipid parameters in a comparison of the 3 liver disease groups.

Table 5. Mean Percentage Changes from the Baseline Lipid Levels (ITT Population)
  • *

    The P value from the between-group comparison was based on an analysis of covariance test. Abbreviations: ITT, intent-to-treat; LDL-C, low-density lipoprotein cholesterol; SD, standard deviation; TC, total cholesterol; TG, triglyceride.

LDL-C (mg/dL)   
 Week 4n = 153n = 154<0.0001
  Mean percentage change (SD)−32.57 (15.142)0.82 (11.945) 
 Week 12n = 155n = 154<0.0001
  Mean percentage change (SD)−30.61 (16.541)2.70 (13.907) 
 Week 36n = 155n = 154<0.0001
  Mean percentage change (SD)−25.85 (19.844)−0.83 (18.994) 
TC (mg/dL)   
 Week 4n = 154n = 155<0.0001
  Mean percentage change (SD)−21.97 (11.091)1.16 (8.488) 
 Week 12n = 156n = 155<0.0001
  Mean percentage change (SD)−20.02 (12.115)3.32 (9.985) 
 Week 36n = 156n = 155<0.0001
  Mean percentage change (SD)−16.71 (14.088)1.28 (13.676) 
TG (mg/dL)   
 Week 4n = 154n = 155<0.0001
  Mean percentage change (SD)−13.75 (27.016)4.80 (31.410) 
 Week 12n = 156n = 155<0.0001
  Mean percentage change (SD)−9.25 (32.435)7.46 (36.438) 
 Week 36n = 156n = 155<0.0005
  Mean percentage change (SD)−6.32 (35.044)8.82 (49.037) 

Safety Evaluations

ALT Elevations.

The cumulative incidence of subjects who met the primary prespecified ALT event rate was lower at all times in the pravastatin group than in the placebo group, but the difference was not statistically significant (Table 6). In general, the mean ALT values were reduced from the baseline in the pravastatin group, whereas the mean ALT and AST were increased from the baseline in the placebo group. None of the subjects in either treatment group had a concomitant total bilirubin value of 2 times the ULN.

Table 6. Cumulative Incidence of Subjects Who Met the Primary Alanine Aminotransferase Safety Endpoint (for All Patient Diagnoses)
WeekPravastatin [n/N (%)]*Placebo [n/N (%)]*95% Confidence IntervalP
  • For pravastatin, n was 163, and for the placebo, n was also 163.

  • *

    The percentages were based on the number of subjects in each treatment group through the specified visit; the subject counts are cumulative.

  • The 95% confidence intervals of the differences in the percentages were based on the asymptomatic risk difference between the treatment groups.

  • The P values were based on Cochran-Mantel-Haenszel tests of the alanine aminotransferase elevation by the treatment, which controlled for 3 liver disease types and the country.

10/160 (0%)1/160 (0.6%)−1.8, 0.60.2850
20/160 (0%)2/160 (1.3%)−3.0, 0.50.1490
42/160 (1.3%)4/160 (2.5%)−4.2, 1.70.3967
66/160 (3.8%)7/160 (4.4%)−5.0, 3.70.7369
86/160 (3.8%)8/160(5.0%)−5.7, 3.20.5468
129/160 (5.6%)11/160 (6.9%)−6.6, 4.10.6095
169/160 (5.6%)12/160 (7.5%)−7.3, 3.50.4805
209/160 (5.6%)15/160 (9.4%)−9.5, 2.00.1897
249/160 (5.6%)17/160 (10.6%)−11.0, 1.00.0951
2810/160 (6.3%)18/160 (11.3%)−11.2, 1.20.1048
3210/160 (6.3%)19/160 (11.9%)−11.9, 0.60.0762
3612/160 (7.5%)20/160 (12.5%)−11.6, 1.60.1379

The proportion of subjects who had sustained elevations in ALT was comparable for both treatment groups [8/160 (5%) and 11/160 (7%) in the pravastatin and placebo groups, respectively]. The time to onset of the primary safety endpoint also did not differ significantly between treatment groups, and there was no statistically significant difference in the percentage of subjects who experienced periods of sustained ALT elevations. Furthermore, when the data were analyzed by baseline ALT entrance criteria (ALT values less than or equal to the ULN and ALT values greater than the ULN), no statistically significant differences were observed for the entire cohort or for the individual subjects with NAFL/NAFLD or chronic hepatitis C (Table 7). Similarly, no statistically significant differences in the ALT event rates were observed between liver disease groups comparing NAFL/NAFLD and chronic hepatitis C either with a normal ALT or an elevated ALT at the baseline.

Table 7. Doubling of ALT in NAFL/NAFLD and Hepatitis C Patients: Normal ALT Versus Elevated ALT at the Baseline
Baseline ALTWeekPravastatinPlaceboP
  1. Abbreviations: ALT, alanine aminotransferase; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease.

NAFL/NAFLD group    
Normal ALT125/45 (11%)3/40 (7.5%)0.6575
 365/45 (11%)4/40 (10%)0.9849
Elevated ALT123/58 (5.2%)4/61 (6.6%)0.7203
 364/58 (6.9%)8/61 (13.1)0.2316
Chronic hepatitis    
Normal ALT120/17 (0%)1/16 (6.3%)0.3173
 360/17 (0%)1/16(6.3%)0.3173
Elevated ALT120/21 (0%)1/27 (3.7%)0.3833
 362/21 (9.5%)5/27 (18.5%)0.3810
All subjects    
Normal ALT125/73 (7.1%)6/68 (8.8%)0.5675
 365/73 (7.1%)7/68 (10.3%)0.3700
Elevated ALT124/87 (4.6%)5/92 (5.4%)0.7363
 367/87 (8%)13/92 (14.1%)0.2098

In order to further evaluate the ALT endpoint, shift tables were constructed to compare the pretreatment baseline values with their respective posttreatment elevations. No statistical differences in ALT doubling were seen between the pravastatin-treated and placebo-treated groups, regardless of the baseline values at any time of the study. Table 8 demonstrates this lack of difference for week 12 (the predefined primary efficacy and safety endpoint) and for week 36 (at the conclusion of the trial).

Table 8. Analysis of ALT Doubling by the Baseline ALT Values: Pravastatin Versus the Placebo
Baseline ALTnPostbaseline Doubling of ALT
  1. Abbreviations: ALT, alanine aminotransferase; ULN, upper limit of normal; X, fold above ULN. For pravastatin versus the placebo at week 12, P = 0.7805; at week 36, P = 0.5871.

Pravastatin at week 12        
<ULN6356 (89%)6(9.5%)1 (1.6%)0000
>1X5714 (24.6%)33 (58%)9 (15.8%)1 (1.8%)000
>2X171 (6%)9 (53%)5 (29.4%)1 (6%)1 (6%)00
>3X501 (20%)2 (40%)2 (40%)000
>4X1000001 (100%)0
>5X100001 (100%)00
Placebo at week 12        
<ULN6151 (83.6%)8 (13.1%)1 (1.6%)001 (1.6%)0
>1X5712 (21.1%)38 (66.7%)7 (12.3%)0000
>2X211 (4.8%)9 (43%)9 (43%)01 (4.8%)1 (4.8%)0
>3X5003 (60%)1 (20%)01 (20%)0
>4X2000002 (100%)0
>5X10001 (100%)000
Pravastatin at week 36        
<ULN5748 (84.2%)8(14%)1 (1.8%)0000
>1X5215 (28.8%)29 (55.8%)5 (9.6%)2 (3.8%)01 (1.9%)0
>2X17011 (64.7%)4 (23.5%)2 (11.8%)000
>3X502 (40%)3 (60%)0000
>4X1000001 (100%)0
>5X10001 (100%)000
Placebo at week 36        
<ULN5241 (78.8%)10 (19.2%)01 (1.9%)000
>1X4816 (33.3%)26 (54.2%)6 (12.5%)0000
>2X173 (17.6%)5 (29.4%)7 (41.2%)2 (11.8%)000
>3X501 (20%)2 (40%)1 (20%)1 (20%)00
>4X2000001 (50%)1 (50%)
>5X1001 (100%)0000


The overall incidence of AEs was similar in the pravastatin group (26.4%) and in the placebo recipients (25.2%), with the most common AE being a headache in both treatment groups. Ten subjects experienced treatment-emergent AEs associated with LFT elevations (6 and 4 in the pravastatin and placebo groups, respectively). There were no patterns or consistent associations between these treatment-emergent AEs associated with changes in the ALT values. No clinically apparent hepatotoxicity was seen in either group.

Twenty-seven subjects discontinued treatment because of AEs (11 and 16 in the pravastatin and placebo groups, respectively). In the pravastatin group, 2 subjects experienced myalgia. In the placebo group, 2 subjects experienced myocardial infarctions, and 1 had rhabdomyolysis. No subject experienced acute exacerbation of underlying liver disease during the trial. One death was reported (in a placebo recipient) due to myocardial infarction that was considered unrelated to the study drug by the investigator.


Pravastatin has been studied in the United States and other countries since 1985, with over 25,000 subjects with primary hypercholesterolemia evaluated in clinical trials; 13,000 of these subjects received pravastatin for periods ranging up to 7 years.37 In the United States, pravastatin (up to 80 mg daily) has been approved and marketed for the reduction of elevated TC and LDL-C levels (and the resulting reduction in cardiovascular disease events) in patients with primary hypercholesterolemia since 1991.38–40

Elevations in LAEs (primarily ALT and AST) are well documented in statin recipients.3, 8, 9 However, in large clinical trials, only approximately 1% of patients receiving a statin with normal baseline ALT levels develop elevations greater than 3 times the ULN, and this is not significantly different from the placebo recipients.5, 7 In a meta-analysis of 49,275 patients compiled by de Denus et al.,5 the proportion of patients having ALT or AST abnormalities was low in both groups (the statin level was 1.14% versus 1.05% for the placebo group; odds ratio = 1.26, 95% confidence interval = 0.99-1.62, P = 0.07). This relatively low rate of transaminitis has also been observed in pravastatin-specific trials. An analysis of 3 prospective studies involving 19,592 patients receiving either pravastatin or a placebo for up to 5.9 years found a similarly low frequency (<1.2%) of ALT greater than 3 times the ULN in both treatment groups.7 These pooled study results from pravastatin clinical trials also included 579 subjects with ALT elevated up to 3 times the ULN at the baseline and demonstrated no excess ALT elevations for pravastatin in comparison with a placebo (5% versus 7.3%, respectively).7 However, no information about the etiology of these elevations was available, and the height of the baseline ALT levels in most subjects was relatively modest (<1.5 times the ULN in most).

For the great majority of statin recipients, these elevations in ALT are asymptomatic, typically do not progress, and usually return to baseline levels, despite the continuation of the drug (so-called tolerance or adaptation).41, 42 Although acute liver failure associated with statins is extremely rare and has not been established as causal in all instances,3, 14, 21 statin safety has been at the center of an active controversy about the treatment of hypercholesterolemia in patients with elevated liver enzymes or overt liver disease.21 This concern is reinforced by the current prescribing information, which warns that all HMG CoA reductase inhibitors have been associated with biochemical abnormalities of liver injury and rare instances of hepatic injury and that either active liver disease or unexplained persistent aminotransferase elevations are contraindications to their use.20, 32

The exact mechanism by which statins cause ALT elevations is uncertain. A dose relationship was demonstrated for lovastatin and appears to exist for the other statins as well.8, 9, 21 Kornbrust et al.4 proposed that the ALT increases and the centrizonal hepatocellular necrosis seen in rabbits might be the result of an exaggerated pharmacologic action that depletes mevalonic acid synthesis, as this form of apparent species-specific hepatotoxicity was prevented (or reversed) in animals given mevalonate along with lovastatin before (or after) hepatotoxicity developed. All of the statins, except for pravastatin, are metabolized by P450 cytochromes,13, 43, 44 although this remains a speculative mechanism for statin-associated LAE elevations. Charles et al.12 described a few instances of positive rechallenge to various statins in a large health maintenance organization database, although features of drug allergy or hypersensitivity are generally absent. A spectrum of histologic and biochemical abnormalities has been reported, with a hepatocellular injury pattern predominating.3 Instances of autoimmune-like injury attributed to statins also have been reported.45–47 Although no apparent correlation exists between the risk of myopathy and hepatotoxicity from statins,48 muscle injury can, at times, cause elevated LAEs that might be confused with drug injury.49

The labeled contraindication to the use of statins in active liver disease seems to be based largely on the assumption that acute hepatotoxicity developing in a patient with impaired liver function would likely be more serious, although evidence-based data to support this theory are lacking.50 Indeed, the late Hyman Zimmerman found little, if any, support in the literature or in his experience for avoiding the use of most potentially hepatotoxic medications in patients with liver disease.33, 51, 52 Although exceptions do exist (such as the increased risk of hepatotoxicity developing in patients with a fatty liver receiving methotrexate or the use of highly active antiretroviral therapy or antituberculosis medications in patients with viral hepatitis B or C),33 statins do not appear to share that susceptibility. Indeed, pravastatin has been used safely in the management of hypercholesterolemia in liver transplant recipients28, 32, 53 and in patients with PBC29, 30 and as a potential treatment option for NAFLD.21, 27 Cirrhosis may alter the pharmacokinetics and metabolism of some statins, although the clinical significance of these findings is unclear.54

Retrospective studies in patients with elevated baseline ALT levels have provided evidence that statins can be used safely in individuals with underlying liver disease.21 Chalasani and colleagues22–24 reviewed a large medical and pharmacy database for information about patients with elevated baseline ALT levels going back to 1987, most having NAFLD.55 Their results indicate that several statins (principally atorvastatin, simvastatin, and lovastatin) did not significantly increase the risk of ALT elevations (up to 10 times the ULN), lead to a concomitant rise in the serum bilirubin levels (above 3 mg/dL), or cause discontinuations among hypercholesterolemic chronic liver disease patients with elevated baseline ALT values receiving a statin in comparison with patients with ALT elevations not on a statin. The Dallas Heart Study also found that statin use was not associated with a greater frequency of ALT elevations in comparison with no statin use.56 A cohort study by Khorashadi et al.25 and a smaller trial by Gibson and Rindone26 did not find any added risk of severe hepatic events in statin-treated patients with chronic hepatitis C.

Together, these studies provide evidence to challenge the notion that statins should be contraindicated in patients with underlying stable, chronic liver disease. Although most authorities opine that the benefits of treating hypercholesterolemia to reduce cardiovascular events more than outweigh the risks of developing serious hepatic injury from a statin,16, 17, 19 the aforementioned studies, all being retrospective in design, leave open a number of questions still to be answered regarding the comparative safety and efficacy of statins in a broad group of chronic liver disease patients treated and followed prospectively.

In an attempt to address several of these remaining questions, this randomized, double-blind, placebo-controlled study was designed to evaluate the lipid-lowering effects and the hepatic and overall safety of 80 mg of pravastatin daily in hypercholesterolemic subjects with chronic, well-compensated liver diseases. All patients in this prospective trial had well-characterized chronic liver disorders of various etiologies (predominantly NAFL/NAFLD and chronic hepatitis C) with ALT elevations on entry up to 5 times the ULN. Although the study was not fully powered to detect small differences in the rate of ALT doubling during treatment, the numbers enrolled were considered adequate for a meaningful safety comparison. Moreover, all pravastatin-treated subjects were treatment-naive prior to statin use, and this eliminated the possibility that we were dealing with a statin-tolerant population. The selection of a doubling of ALT (rather than a 3-fold or higher elevation) was considered the simplest and most direct means of detecting a difference between the groups.

The results demonstrate that high-dose pravastatin is efficacious and safe in this patient population. Overall, there was an 8% rate (12/160 subjects) of ALT elevation events (defined as a doubling of baseline values) in the pravastatin group, which was lower at all times over 36 weeks (although not statistically significantly different from) than the 13% rate (20/160 subjects) in the placebo group. There were still no differences observed when patients with a normal baseline ALT were separated from those with an elevated ALT at the baseline. Although about two-thirds of our patients had NAFL or NAFLD and about one-quarter had chronic hepatitis C, when a comparison of ALT values was made between these 2 main liver disease groups and with the pooled group of other chronic liver diseases, no differences in the efficacy or ALT events were seen.

The finding that ALT events were numerically more likely to occur in the placebo group in this trial is similar to what was observed by Chalasani and colleagues.22–24 In their retrospective study using primarily simvastatin and atorvastatin, ALT elevations greater than 10-fold (corresponding to a doubling of a baseline value of 5 times the ULN in our study) were recorded in 4.7% of statin-treated patients with elevated baseline ALT values versus 6.4% of those with a history of elevated ALT not on a statin.22 In their second analysis examining the effects of lovastatin on liver enzymes, ALT elevations were seen in 6.6% of patients on lovastatin versus 11% in the cohort not receiving a statin.23 The baseline levels of TC, LDL-C, HDL, and TGs in these patients were also similar to those in our trial, and this raises the possibility that treatment with a statin in these patients (most of whom had NAFLD) might have produced a beneficial, enzyme-lowering effect on the underlying steatohepatitis. Although a small pilot study of pravastatin27 suggests that this may be the case, neither our study nor those of Chalasani and colleagues22–24 were designed or powered to be able to show a specific treatment effect of statins on the underlying liver disease.

Because no significant ALT elevations or other hepatobiliary AEs occurred among our patients with liver disease receiving high-dose pravastatin with respect to the placebo group, we believe that this prospective study provides the reassurance required to consider initiating pravastatin therapy for patients with chronic, stable liver disease and hypercholesterolemia so as not to limit their access to the primary and secondary cardiovascular disease prevention benefits of cholesterol reduction.16, 17, 19, 31, 56–63

In conclusion, high-dose pravastatin administered daily to hypercholesterolemic subjects with chronic, well-compensated liver disease in this randomized, placebo-controlled trial was both efficacious and safe. Pravastatin significantly lowered TC, LDL-C, and TGs in comparison with the placebo. Although not powered to detect small differences in ALT events between the groups, the study did demonstrate that the frequency of ALT elevations that doubled from either a normal or elevated baseline in the pravastatin-treated group was not statistically significantly different from that of placebo-treated patients. As a result, this prospective trial confirms the findings of several retrospective studies demonstrating the lack of significant hepatotoxicity developing in patients with compensated chronic liver disease treated with a statin and supports the recommendation that hypercholesterolemic patients with compensated chronic active liver disease should not be denied access to pravastatin (or similar agents) solely on the basis of their liver disease.


The authors thank Eileen Kelly of Bristol-Myers Squibb for her invaluable assistance with the manuscript preparation.


Clinical Investigators

The clinical investigators were as follows: Jeffrey R. Medoff, M.D., Greensboro, NC (principal investigator); Nuzhat Abbassi, M.D., Miami, FL; Nezam Afdahl, M.D., Boston, MA; Lawrence Alder, M.D., San Antonio, TX; Melchor Alpizar, M.D., Mexico City, Mexico; Nancy Bach, M.D., New York, NY; Charles Barish, M.D., Raleigh, NC; Stephen Brady, D.O., Naples, FL; Michael Bennett, M.D., San Diego, CA; Jay Bernstein, D.O., Phoenix, AZ; Herbert Bonkovsky, M.D., Farmington, CT; Thomas Boyer, M.D., Tucson, AZ; Osvaldo Brusco, M.D., Corpus Christi, TX; Marie Buscemi, M.D., Eugene, OR; Michael Mirhej, M.D., Eugene, OR; Natalie Bzowej, San Francisco, CA; Stephen Caldwell, M.D., Charlottesville, VA; Travis Calhoun, M.D., Clarksville, TN; Tommy Carmen, M.D., Clarksville, TN; Jaime Carranza, M.D., Morelia, Mexico; Richard A. Cazen, M.D., San Francisco, CA; Kimy Charani, M.D., Oro Valley, AZ; Richard Chasen, M.D., Laurel, MD; Ramsey Cheung, Palo Alto, CA; Henry Ciozcek, M.D., Brooklyn, NY; Clinton Corder, M.D., Oklahoma City, OK; Thomas Dallman, M.D., Bullhead City, AZ; Michale Demicco, M.D., Oklahoma City, OK; Carlos Dominguez-Reyes, M.D., Aguascalientes, Mexico; William Espinoza, M.D., Miami, FL; Michael Farrell, M.D., Richmond, VA; Fred Fowler, M.D., Charlotte, NC; Bradley Freilich, M.D., Kansas City, MO; Reem Ghalib, M.D., Dallas, TX; Robert Hardi, M.D., Chevy Chase, MD; Willian Harlan III, M.D., Asheville, NC; Douglas Heuman, M.D., Richmond, VA; Adesh Jain, M.D., Slidell, LA; Vikram Jayanty, M.D., Houston, TX; Albert Johary, M.D., Dunwoody, GA; Barry Kaufman, M.D., Egg Harbor Township, NJ; Paul King, M.D., Columbia, MO; Mariusz Klin, M.D., Panama City, FL; Howard Knapp, M.D., Billings, MT; Whitfield Knapple, M.D., North Little Rock, AR; Valli Kodali, M.D., Fayetteville, NC; Venkates Kolli, M.D., El Paso, TX; Larry Levine, M.D., Spring Hill, FL; Thomas Littlejohn, M.D., Winston-Salem, NC; K. Jean Lucas, M.D., Morehead City, NC; Amer Malik, M.D., North Providence, RI; Thomas Marbury, M.D., Orlando, FL; Paul Martin, M.D., Los Angeles, CA; Manuel Mendoza, M.D., Bellflower, CA; Dennis Mikolich, M.D., Cranston, RI; Albert Min, M.D., New York, NY; Martin Mollen, M.D., Phoenix, AZ; Linda Munoz, M.D., Monterrey, Mexico; Mark Murphy, M.D., Savannah, GA; James Nelson, M.D., Milwaukee, WI; Michael Schmatz, M.D., Milwaukee, WI; Tuan Nguyen, M.D., San Diego, CA; Steven O'Marro, M.D., Springfield, IL; Douglas Leigh, M.D., Springfield, IL; Alberto Pasquetti, M.D., Mexico City, Mexico; Michael Perley, M.D., Lakewood, CA; Patrick Peters, Jr., M.D., San Antonio, TX; Jorge Poo, M.D., Mexico City, Mexico; John Pullman, M.D., Butte, MT; Saman Ratnayke, M.D., Bakersfield, CA; Nancy Reau, M.D., Chicago, IL; Robert Reindollar, M.D., Charlotte, NC; Maribel Rodriguez Torres, M.D., Santurce, Puerto Rico; Angel Rosario, M.D., New Port Richey, FL; Dale Rosenberg, M.D., Lancaster, PA; Raymond Rubin, M.D., Atlanta, GA; Vinod Rustgi, M.D., Fairfax, VA; Michael Ryan, M.D., Norfolk, VA; Alan Safdi, M.D., Cincinnati, OH; Julio Salcedo, M.D., Washington, DC; Joseph Saponaro, M.D., Jupiter, FL; Leonardo Sauque Reyna, M.D., Cuernavaca, Mexico; Sherwyn Schwartz, M.D., San Antonio, TX; Gregory Serfer, D.O., Hollywood, FL; Tariq Shakoor, M.D., Cincinnati, OH; Nav Grandhi, M.D., Cincinnati, OH; Kenneth Sherman, M.D., Cincinnati, OH; Mitchell Shiffman, M.D., Richmond, VA; Mark Stern, M.D., Decatur, GA; James Sullivan, M.D., Birmingham, AL; Paul Thuluvath, M.D., Baltimore, MD; Thomas Werth, M.D., Charlotte, NC; Ira Willner, M.D., Charleston, SC; Robert Wohlman, Bellevue, WA; Lawrence Wruble, M.D., Memphis, TN; and Zobair Younossi, M.D., Falls Church, VA.

The members of the Independent Hepatic Safety Board were Eugene R. Schiff, M.D., Miami, FL, and Steven Schenker, M.D., San Antonio, TX.